ByteLite LLC

Research program

Exact, deterministic data representation.

Representing data so it rebuilds exactly, byte for byte, every time, with every byte accounted for.

Proven in internal tests

What works today.

  • 4,515 / 4,515

    byte-exact rebuilds

    Every test encoding rebuilt exactly.

    Each one was decoded and matched to its original byte for byte before it was accepted. Separate 8 MiB and 64 MiB files also rebuilt exactly.

    Internal certification run, one machine, October 2026. The test set is small, and the representation is still larger than its input.

  • 83,332

    checks passed, zero failed

    Same input, same output.

    34 of 34 test suites and every determinism gate passed in release and debug builds.

    Internal certification run, one machine, October 2026.

  • 4.36 to 4.81 MB

    peak memory, 1 KiB to 64 MiB

    Memory stays flat as files grow.

    Peak memory stayed nearly constant across inputs from 1 KiB to 64 MiB.

    Measured on one machine.

Net size reduction has not been demonstrated. The current development representation is larger than its input, and no compression ratio is claimed.

The rule every result must pass

Exact, or it does not count.

  1. Declare Record the input and its size before anything runs.
  2. Represent Count everything needed to rebuild it. Nothing held outside.
  3. Rebuild Decode and compare with the original, byte for byte.
  4. Report Publish a size only if the rebuild was exact.
Conceptual illustration How results are checked, not how a representation is produced. The methods are proprietary.

Deterministic, not predictive

The core encoding and decoding process is deterministic. It does not rely on probabilistic inference or confidence scores to decide what the original data probably was.

A mathematical limit we respect

No lossless method can make every possible input smaller. There are more possible inputs of a given length than there are shorter outputs to hold them, so some inputs must stay the same size or grow. ByteLite LLC therefore never claims universal reduction. Any reduction it reports will be measured on a declared input, using the complete representation.

Earlier versions of this website described universal reduction as the goal. That was wrong and has been withdrawn.

Milestones

Every gate, and where it stands.

  • Exact reconstruction

    The original data is rebuilt byte for byte from its representation.

    Evidence and next gate

    Shown: Every case in the internal test set is decoded and compared with the original byte for byte before the representation is accepted.

    Not yet shown: Exact reconstruction verified by anyone outside ByteLite LLC. Detection of corruption that happens to a representation after it was created.

    Next gate: The same check repeated on a broad public test collection.

  • Complete artifact accounting

    Every size figure should count the complete representation, including everything needed to rebuild the original.

    Evidence and next gate

    Shown: For the current single-round format, reported sizes are measured from the complete artifact file.

    Not yet shown: Complete accounting for every output form. Independent audit of the accounting.

    Next gate: Accounting rules published in a form an outside reviewer can check.

  • Deterministic operation

    The same input and the same software always produce the same representation.

    Evidence and next gate

    Shown: Repeated runs on one machine produce identical output in internal tests.

    Not yet shown: Identical output across separate machines. Independent confirmation.

    Next gate: Determinism confirmed across separate machines by an outside party.

  • Measured net size reduction

    A complete, self-contained representation that is smaller than the original for real inputs, measured and reported per input.

    Evidence and next gate

    Shown: Not demonstrated. The current development representation is larger than its input.

    Not yet shown: Net reduction on any input class. Any compression ratio.

    Next gate: A measured, complete representation smaller than its input on a declared test collection.

  • Independent evaluation and production qualification

    Reproduction of results by an outside party, followed by qualification for real workloads.

    Evidence and next gate

    Shown: Not demonstrated.

    Not yet shown: Independent reproduction. Production qualification.

    Next gate: An agreed evaluation protocol with an outside reviewer.

Check a result yourself.

Nothing is available to license yet. Proposals to evaluate a specific result are welcome.

Propose an evaluation Licensing models